DURABILITY ANDDAMAGE TOLERANCE OFALUMINUM CASTINGS
M.W. OZELTON DTIOG.R. TURK A kELE CTE *T
OCT 0 5 1988NORTHROP CORPORATIONAIRCRAFT DIVISIONONE NORTHROP AVENUE DHAWTHORNE, CALIFORNIA 90250
September 1988NOR 88-23
Second Interim Technical RerortFor Period June 1987 - May 1988Contract No. F33615-85-C-5015
Approv3d For Public Release: Distributior, Unlimited
Prepared For:MATERIALS LAL.JRATORYAIR FORCE WRIGHT AERONAUTICAL LABORATORIESWRIGHT-PATTERSON AFB, OHIO 45433-6533
88 10 4 10(
UNCLASSI FIED
SECURITY CLASSIFICATION OF THIS PAGE
REPORT DOCUMENTATION PAGEIa. REPORT SECURITY CLASSIFICATION lb. RESTRICTIVE MARKINGS
Unclassified2a SECURITY CLASSIFICATION AUTHORITY 3. DISTRIBUTION/AVAILABILITY OF REPORT
2b. DECLASSIFICATION /DOWNGRADING SCHEDULE Unlimited
4. PERFORMING ORGANIZATION REPORT NUMBER(S) 5. MONITORING ORGANIZATION REPORT NUMBER(S)
NOR 88-23
6a. NAME OF PERFORMING ORGANIZATION 6b OFFICE SYMBOL 7a NAME OF MONITORING ORGANIZATION
Northrop Corporation (if applicable) Air Force Wright AeronauticalAircraft Division 3872/62 Laboratories
6c. ADDRESS (City, State, and ZIP Code) 7b ADDRESS (City, State, and ZIP Code)
One Northrop Avenue AFWAL/MLSEHawthorne, CA 90250 Wright-Patterson AFB, OH 45433
8a. NAME OF FUNDING/SPONSORING 8b OFFICE SYMBOL 9 PROCUREMENT INSTRUMENT IDENTIFICATION NUMBERORGANIZATION Air Force Wright (if applicable) F33615-85-C-5015Aeronautical Laboratories
BC. ADDRESS (City, State, and ZIP Code) 10 SOURCE OF FUNDING NUMBERSPROGRAM PROECT ITASK IWORK UNIT
Wright-Patterson AFB, OH 45433 ELEMENT NO NO NO ACCESSION NO
11 TITLE (Include Security Classification)
Durability and Damage Tolerance of Aluminum Castings (Unclassified)
12. PERSONAL AUTHOR(S) M. W. Ozelton and G. R. Turk
13a. TYPE OF REPORT 113b. TIME COVERED 114. DATE OF REPORT (Year, Month, Day) IS PAGE COUNT
Interim FROM 6/1/87 TO 5/31/88 1988, September 7616 SUPPLEMENTARY NOTATION
17 COSATI CODES 18 SUBJECT TERMS (Continue on reverse if necessary and identify by block number)
FIELD GROUP SUB-GROP ....... Aluminum castings, -93i,-82Ol, mechanical properties,durability, damage tolerance, equivalent initial flawsize, fatigue, fracture toughness. --i i--
19 ABSTRACT (Continue on reverse if necessary and identify by block number)
A durability and damage tolerance property data base for D357-T6 and B201-T7 cast bythree commercial foundries was developed. The specifications for the two alloys wereselected based on the results of screening evaluations which are included in the firstinterim report published in September 1987.
Tensile, fatigue, and fracture toughness properties were determined and microstructuraland fractographic evaluations were conducted. The fatigue data (stress-life and crackgrowth rate) were used to calculate an equivalent initial flaw size for both alloys andwill be used to assess the applicability of current DADT specifications to castings.
20 DISTRIBUTION /AVAILABLITY OF ABSTRACT 21 ABSTRACT SECURITY CLASSIFICATION
E-UNCLASSIFIED/UNLIMITED [] SAME AS RPT l DTIC USERS Unclassified22a NAME OF RESPONSIBLE INDIVIDUAL 22b TELEPHONE (Include Area Code) 22c OFFICE SYMBOL
DO FORM 1473, 84 MAR 83 APR edition may be used until exhausted SECURITY CLASSIFICATION OF THIS PAGEAll other editions are obsolete Unc lass i fi ed
f-i
ABSTRACT
A durability and damage tolerance property data base was
developed for D357-T6 and B201-T7 alloys that were cast by three
commercial foundries. The specifications for the two alloys were
selected based on the results of screening evaluations which are
included in the first interim report published in September,
1987.
Tensile, fatigue, and fractu-e toughness properties were
determined and microstructural and fractographic evaluations were
conducted. Fatigue data (stress-life and crack growth rate) for
both alloys were used to calculate an equivalent initial flaw
size, which will be used to assess the applicability of current
DADT specifications to castings.
%0
LDTIC TAB
*P ; '.. .,x , ,0-(N€, C r
S.. .. . ... ..........--------
-~ I
iii
PREFACE
This second interim report covers the work performed under
Contract F33615-85-C-5015 from June 1987 to May 1988 by Northrop
Corporation, Aircraft Division, Hawthorne, California, under
Project Number 2418. The program is administered under the
technical direction of Mary Ann Phillips, AFWAL/MLSE, Wright-
Patterson AFB, Ohio 45433.
The work is being performed by Northrop's Metallic Materials
and Processes Research and Development Department. Dr. M. W.
Ozelton is the Program Manager and Mr. G. R. Turk is the
Principal Investigator. The contributions of the following
people are acknowledged: P. G. Porter, equivalent initial flaw
size analyses; K. J. Oswalt for many invaluable discussions; S.
Hsu, mechanical testing; D. R. Drott, H. E. Langman, and M. A.
Arrieta for coordinating and carrying out the many and varied
activities to obtain the test data; and J. F. Williams for
preparing the manuscript.
The major subcontractors for the program are Alcoa, Premium
Castings Division, Corona, California; Hitchcock Industries Inc.,
Minneapolis, Minnesota; Fansteel Wellman Dynamics, Creston, Iowa;
and Dr. L. Adler of Ohio State University (NDI consultant).
v
CONTENTS
SECTION PAGE
1 INTRODUCTION.......................................... 1
2 OBJECTIVE............................................. 3
3 PROGRAM APPROACH...................................... 5
4 EXPERIMENTAL PROCEDURES............................... 9
4.1 PRODUCTION OF CASTINGS........................... 9
4 .2 TEST PROCEDURES................................. 20
4.3 EQUIVALENT INITIAL FLAW SIZE ANALYSIS ..........22
5 RESULTS AND DISCUSSION............................... 27
5.1 D357-T6......................................... 27
5.1.1 Composition.............................. 285.1.2 Microstructure........................... 285.1.3 Tensile Properties...................... 355.1.4 Fatigue Properties....................... 375.1.5 Fracture Toughness...................... 395.1.6 Equivalent Initial Flaw Size ............43
5.2 B201-T7......................................... 50
5.2.1 Composition.............................. 505.2.2 Microstructure........................... 505.2.3 Tensile Properties...................... 555.2.4 Fatigue Properties....................... 575.2.5 Fracture Toughness....................... 605.2.6 Stress Corrosion......................... 605.2.7 Equivalent Initial Flaw Size ............63
6 CONCLUSIONS.......................................... 67
7 REFERENCES........................................... 69
vii
LIST OF FIGURES
FIGURE PAGE
1 Program Outline ........................................ 6
2 Rigging System Used by Foundry A to ProduceD357 Plates ........................................... 13
3 Rigging System Used by Foundry B to ProduceD357 Plates ........................................... 15
4 Rigging System Used by Foundry C to Produce
B201 Plates ........................................... 19
5 Typical D357-T6 Microstructure ......................... 30
6 Backscatter SEM Micrograph of D357-T6 ................. 33
7 TEM Micrograph of D357-T6 Showing Si Particles ........ 33
8 TEM Micrograph of D357-T6 Showing a Needle-ShapedPhase Containing Ti ................................... 33
9 (a) TEM Micrograph of D357-T6 Showing a Needle-Shaped Particle Containing Ti and Smaller Mg 2 SiParticles ............................................. 34
(b) Selected Area Diffraction Pattern of the NeedlePhase of Figure 9(a) .................................. 34
10 D357-T6 Stress-Life Fatigue Data ....................... 38 S
11 D357-T6 Strain-Life Fatigue Data ....................... 38
12 D357-T6 Fatigue Crack Growth Rate Data ................ 40
13 Typical Crack Initiation Sites in D357-T6 FatigueSpecimens ............................................. 41
14 Relationship Between Fracture Toughness andYield Strength for D357-T6 ............................ 44
15 Weibull Distribution of D357-T6 EIFS Data ............. 47
16 Measured Area of Defect Size in D357-T6 ............... 49
17 Typical B201-T7 Microstructure ......................... 51
18 TEM Micrograph of B201-T7 Showing the Theta Phase ..... 53 5
19 Backscatter SEM Image of B201-T7 ....................... 54
ix S
LIST OF FIGURES
FIGURE PAGE
20 STEM Micrograph of B201-T7 Showing a Phase Rich inAl, Cu, Fe, and Mn .................................... 54
21 TEM Micrograph of B201-T7 Showing a Phase Rich inAl, Cu, and Mn ........................................ 54
22 TEM Micrograph of B201-T7 Showing a Phase Rich in
Al and Ti ............................................. 54
23 B201-T7 Stress-Life Fatigue Data ....................... 58
24 B201-T7 Strain-Life Fatigue Data ....................... 58
25 B201-T7 Fatigue Crack Growth Rate Data ................ 59
26 Foreign Material on a B201-T7 Fatigue SpecimenFracture Surface ...................................... 61
27 Relationship Between Fracture Toughness and YieldStrength of B201-T7 ................................... 62
28 Weibull Distribution of B201-T7 EIFS Data ............. 65
x
LIST OF TABLES
TABLE PAGE
1 Composition Specification for the D357-T6Verification Plates ................................... 11
2 Composition Specification for the B201-T7
Verification Plates ................................... 16
3 Verification Tests for D357-T6 and B201-T7 ............ 21
4 D357 Verification Material Melt Compositions .......... 29
5 D357 DAS and Si Particle Data ......................... 31
6 Tensile Properties of 1.25-inch-thick, Water- andGlycol-quenched D357-T6 Plates ........................ 36
7 Fatigue Crack Initiation Sites for D357-T6 FatigueSpecimens ................................................................. 40
8 D357-T6 Fracture Toughness Data ....................... 42
9 D357-T6 Equivalent Initial Flaw Size Data ............. 45
10 Comparison of Predicted and Measured EquivalentInitial Flaw Size for D357-T6 ......................... 48
11 B201 Verification Material Melt Compositions .......... 51
12 Tensile Properties of 1.25-inch-thick
B201-T7 Plates ........................................ 56
13 B201-T7 Fracture Toughness Data ....................... 62
14 B201-T7 Equivalent Initial Flaw Size Data ............. 64
15 Comparison of Predicted and Measured EquivalentInitial Flaw Size for B201-T7 ......................... 66
6 0
xi0
SECTION 1
INTRODUCTION
The use of castings for aircraft applications offers the
potential for significant cost and weight savings compared with
built-up structures. Recurring cost reductions greater than 30
percent are possible through reduced machining and assembly
costs. Weight savings can be achieved, for example, by
eliminating lap joints and fasteners. A reduction in fastener
holes decreases the number of potential flaw initiation sites.
Castings also allow greater shape flexibility compared with
built-up structures.
Castings are not currently used in fatigue-critical aircraft
structure because of the inability to accurately inspect thick
sections, and a lack of durability and damage tolerance (DADT)
data and design allowables that take into account the effects of
inherent casting discontinuities and microstructural features on
DADT properties. The absence of the required DADT data for
castings precludes compliance with the current military DADT
specifications, MIL-A-83444 and MIL-A-87221. As a result, it is
not possible to take advantage of the inherent cost and weight
savings of castings for fatigue-critical aircraft applications.
The goals of the "Durability and Damage Tolerance of
Aluminum Castings" (DADTAC) program are to (a) investigate a
nondestructive inspection (NDI) technique that has the potential
for providing a quantitative assessment of the discontinuities
present in aluminum castings, (b) identify the process variables
and discontinuities that influence the DADT properties, and (c)
characterize the DADT properties of the aluminum casting alloys
A357 and A201. On completion of the program, revisions to
material, process, and DADT specifications will be recommended,
if necessary.
.. .
The first interim report [1] covered work conducted during
the first 21 months of the program under Tasks 1 and 2 of Phase
I. NDI investigations and the effects of process variables on
the properties of A357-T6 and A201-T7 castings were described.
This second interim report covers additional work conducted
during the following 12 months under Task 2 of Phase I (see
Section 3 for further details). Tensile, microstructural, and
durability and damage tolerance properties of A357-T6 and A201-T7
produced using the specifications selected earlier [1] in Task 2
were determined. These alloys were produced according to AMS
specifications 4241 [2] and 4242 [3], respectively, though a
requirement for a silicon modifier was added to AMS 4241.
Therefore, A357 and A201 will be referred to as D357 and B201,
respectively, throughout this report, per the AMS specifications.
2
SECTION 2
OBJECTIVE
The overall objective of the DADTAC program is to generate
durability and damage tolerance data for premium quality aluminum
castings and advance foundry technology to expedite the use of
these castings in primary aircraft structures. In addition,
recommendations for modifying material, process, and DADT
sb
! spcifiatins wll b mae asrequre3
SECTION 3
PROGRAM APPROACH
The DADTAC program, which consists of three phases, is
summarized in Figure 1. A brief description of the technical
activities associated with each phase is provided below.
Phase I - DADT of Aluminum Castings Quality Assessment
Task 1 - NDI Assessment State-of-the-art NDI methods were
evaluated using the expertise of the program NDI consultant, Dr.
L. Adler of Ohio State University, to select a new NDI technique
for determining the soundness of the cast plates and components
tested during the program. Precise knowledge of the size,
location, and types of discontinuities in the test castings was
essential so that detected defects could be related to
radiographic soundness.
Task 2 - Alloy Definition Concurrent with Task 1, the
effects of process variables on the durability and damage
tolerance (DADT) properties of D357-T6 and B201-T7 were assessed.
Screening tests were conducted during the Screening portion of
this task to identify the process variables that provided the
optimum balance of tensile and DADT properties for each of the
two alloys. Under the Verification portion of this task
additional castings were then made using those process variables
that provided the optimum properties, and the DADT properties
were comprehensively characterized. This verification material
will become the baseline for the remainder of the program.
Task 3 - Effects of Defects on DADT Properties The effects
of discontinuities and metallurgical features (microstructure) on
the DADT properties of D357-T6 are being determined. Because the
emphasis of this program is on D357-T6, only the effect of
microstructure will be determined for B201-T7.
5
SDADTAC
Phase I - DADTAC Phase 11I- DADTAC Phase III - DADTAC
Quality Assessment Quality Verification Data Consolidation
" NDI Assessment • Make Aircraft • Materials andCastings to Process Specification• Alloy Definition Specification
- Screening S Recommend- Verification • Validate Phase I Char ges to Existing
" Effect of Defects Data DADT Specifications
on DADT Properties
• Applicability ofExisting DADTSpecifications
Figure 1. Program Outline
6
Task 4 - Applicability of Existing DADT Specifications The
applicability of the MIL-A-83444 and MIL-A-87221 specifications
to D357-T6 and B201-T7 aluminum castings will be assessed, based
on the data generated in Tasks 1-3.
Phase II - DADT of Aluminum Castings Quality Verification
During Phase II, the DADT properties obtained during the
Phase I, Task 2 verification testing will be confirmed by testing
coupons removed from a cast aircraft component. After
considering several existing castings, one that meets the program
requirements will be chosen. Four castings of the selected
design will be produced based on the processing and material
specification guidelines from Phase I, and will then be subjected
to NDI and DADT property evaluations. The results will be
compared with those obtained during Phase I to ensure that the
process requirements can be scaled-up from small cast plates to a
large aircraft casting.
Phase III - DADTAC Program Data Consolidation
In Phase III, the results of Phases I and II will be
consolidated. Material and process specifications, including any
necessary silicon modification and NDI requirements, will be
prepared for use in producing premium quality aluminum castingsfor primary aircraft structure. Necessary modifications to
existing damage tolerance specifications, such as MIL-A-83444 andMIL-A-87221, will be identified and pursued as required.
7
SECTION 4
EXPERIMENTAL PROCEDURES
4.1 PRODUCTION OF CASTINGS
D357-T6 and B201-T7 plates were cast by Hitchcock
Industries, Fansteel Wellman Dynamics, and Alcoa according to
specifications derived from the screening evaluations reported in
the first interim report [1]. The specifications were based on
the process variables that provided the best overall balance oftensile and DADT properties. The use of three foundries to
produce material to the same specification enabled the property
variations typical of the casting industry to be determined. The
D357-T6 results for one of the three foundries were not available
in time for inclusion in this report. The foundries are referred
to as Foundries A, B, or C, and do not correspond to the order
listed above.
Plates that were 16 inches x 6 inches x 1.25-inch-thick were
used for the verification evaluations of Task 2 covered in this
report. This plate thickness enabled fracture toughness speci-
mens of a valid thickness to be obtained. However, a limited
number of plates of the same thickness used previously in the
screening evaluations of Task 2 [1] were also made and evaluated
to (a) assure that the verification material properties were
similar to those obtained from the best screening task plates and
(b) to obtain an indication of the effect of thickness on tensile
properties.
The specifications given to the foundries, which were
structured to be similar to the AMS 4241 [2] and 4242 [3]
aluminum casting alloy specifications for D357-T6 and B201-T7,
respectively, are outlined in the following sections. Key
mechanical and microstructural properties were specified, but the
foundries were given the freedom to select the production methods
that would best enable them to produce plates with the required
properties. The plates were designated in all areas.
9 0
4.1.1 D357-T6
Screening evaluation data [1] indicated that there was no
conflict in the requirements for optimizing tensile and DADT
properties. Therefore, the AMS 4241 composition specification
[2] was used, Table 1, with the addition of strontium as a
silicon modifier, which was shown in the screening evaluations
[1] to improve mechanical properties, particularly the NTS/YS
ratio (an indicator of toughness) and ductility. Although Sr was
used during the screening evaluation, one foundry preferred to
use sodium for the verification plates because previous
experience indicated that it provided benefits at the foundry
without impairing the modification process. The other foundry
used Sr.
A maximum DAS of 0.0024 inch throughout the plates was
specified. This DAS value was selected based on the results of
the screening evaluations; plates with a more rapid
solidification rate (DAS of 0.0024 inch or less) had higher
ultimate tensile strengths (UTS), ductilities, notched tensile
strengths, NTS/YS ratios, and fatigue lives than those with a
slower cooling rate (larger average DAS).
Each foundry produced two plates of the same size as the
screening plates (12 inches x 6 inches x 0.75 inch) and eight
larger plates (16 inches x 6 inches x 1.25 inch). The smaller
plates, which were the same size as those used in the screening
evaluation portion of Task 2, were tested to assure that the
properties of the optimally-produced screening material could be
reproduced using the verification material specifications.
Because castings are often quenched in a glycol/water solution to
reduce distortion (by reducing the cooling rate) of complex parts
and/or those that have significant thickness variations, three
of the eight large plates were quenched in a room temperature
glycol (25%)/water solution (GQ); the remaining plates werequenched in room temperature water (WQ). The rigging system and
10
TABLE 1. COMPOSITION SPECIFICATION(l) FOR THED357 VERIFICATION PLATES
Element Range (wt %)
Min. Max.
Silicon 6.5 7.5
Magnesium 0.55 0.6
Titanium 0.10 0.20
Beryllium 0.04 0.07
Strontium 0.008 0.014(2)Sodium 0.002 0.012
Iron - 0.12
Manganese - 0.10
Others, each - 0.05
Others, total - 0.15
Aluminum Balance
()AMS 4241 plus Si modifier
(2 )Sr or Na was used, not both.
0 ii
0
rI
aging treatments were the same for the GQ and WQ plates, allowing
the effect of quench medium on DADT properties to be determined.
The specified minimum tensile properties were based on the
results obtained during the screening evaluations and were iden-
tical to the AMS 4241 specification for designated areas, as
follows:
Ultimate tensile strength - 50 ksi
Yield strength at 0.2% offset - 40 ksi
Elongation - 3%
The foundries selected their preferred aging procedures to meet
the mechanical property requirements.
A flat vacuum density test result was specified to assure
that the melt hydrogen gas content, which influences the amount
of porosity in the final casting, was minimized.
No weld repair was allowed, and each plate was required to
be Grade B or better by 1% sensitivity radiography according to
MIL-STD-2175.
4.1.1.1 D357 Castings
The rigging system for the 0.75-inch-thick plates was
described in an earlier report [1]. The rigging systems and the
heat treatment procedures used by Foundries A and B for the 1.25-
inch-thick D357-T6 plates are described below.
Foundry A
The rigging, shown in Figure 2, consisted of a tapered down
sprue leading into a pouring well and runner system which fed the
vertical mold cavity. Three 1.25-inch-thick copper chills were
12
Figure 2. Rigqing System Used by Foundry Ato Produce D357 Plates.
13
used, and a fiberglass screen was placed between the pouring well
and runner to trap foreign material.
The plates were solution heat treated for 15 hours at 1015F
in a drop bottom furnace. The temperature was -ontrolled to
within +5F and -10F. The plates were quenched .n room
temperature water or a 25% glycol-water solution, and were
artificially aged at 335+5F for 6 hours.
Foundry B
The rigging for the plates produced by Foundry B is shown in
Figure 3. Eight 2-inch by 2-inch by 1.5-inch-thick iron chills
placed end on end were used in the center of the horizontal cope,
and one 3-inch by 3-inch by 16-inch-long copper chill was used in
the drag to obtain directional and rapid solidification. Eight
insulated risers spaced at approximately 4-inch intervals were
used to feed the casting during solidification. The mold cavity
was gravity fed through an insulated gating system which con-
tained ceramic filters to trap foreign material.
The plates were solution heat treated at 1010+5F for 24
hours and quenched in room temperature water or a 25% glycol-
water solution. The plates were then artificially aged at 315+5F
for 13.5 hours.
4.1.2 B201-T7
Similar to D357-T6, the screening evaluation data showed
that there was no conflict in the requirements for optimizing
tensile and DADT properties. Consequently, the composition
specified in AMS 4242 (Table 2) [3] was used for the verification
plates. Each foundry produced two 0.75-inch-thick plates and
five 1.25-inch-thick plates and selected their preferred aging
treatment to meet the mechanical property requirements. All
areas of the plates required designated properties and all of the
14
IfI
a. Cope Side of Casting.
!0
b. Runner and Filter System.
Figure 3. Rigging System Used by Foundry B toProduce D357 Plates.
15
m m m m m
TABLE 2. COMPOSITION SPECIFICATION(1 ) FOR THEB201 VERIFICATION PLATES.
Element Range (wt %)
Min. Max.
Copper 4.50 5.0
Silver 0.40 0.8
Maganese 0.20 0.50
Magnesium 0.20 0.30
Titanium 0.15 0.35
Iron - 0.05
Silicon - 0.05
Others, each - 0.05
Others, total - 0.15
Aluminum Balance
()AMS 4242
16
plates were quenched in room temperature water. A flat vacuum
density test result was required. To minimize microshrinkage,
which may have a deleterious effect on fatigue life, all B201
plates were HIPed. HIPing consisted of step-heating the as-cast
plates to 950F and holding for 3 hours in an inert gas atmosphere
at 15,000 psi.
The specified minimum tensile properties were based on the
results obtained during the screening evaluations and were iden-
tical to the AMS 4242 specification requirements for designated
areas, as follows:
Ultimate tensile strength - 60 ksi
Yield strength at 0.2% offset - 50 ksi
Elongation - 3%
Similar to D357-T6, weld repairs were not allowed, and each
plate was required to be of a Grade B or better radiographic
quality (1% sensitivity).
4.1.2.1 B201 Castings
Similar to the D357 plates, each foundry designed its own
rigging system and chose the solution heat treat and aging param-
eters they thought best to achieve the T7 temper with the
required mechanical properties. The rigging system for the 0.75-
inch-thick plates was described in an earlier report [1]. The
rigging systems and heat treatment procedures used by Foundries
A, B, and C for the 1.25-inch-thick plates are described below.
Foundry A
The rigging for plates produced by Foundry A was the same as
that used for the D357 plates (Figure 2).
17
The plates were loaded in a furnace below 50OF and solution
heat treated as follows:
2 hours at 940+1OF
2 hours at 960+1OF
16 hours at 980+1OF
The plates were then quenched in room temperature water and
artificially aged for 5 hours at 380+5F.
Foundry B
The rigging for the plates produced by Foundry B was similar
to that shown in Figure 3 except that no chills were used in the
cope. One 2-inch by 3-inch by 16-inch-long copper chill was used
in the drag to obtain directional solidification. Eight
insulated risers spaced at approximately 4-inch intervals were
used to feed the casting during solidification. The mold cavity
was gravity fed through an insulated gating system which
contained ceramic filters to trap foreign material.
The plates were loaded into a furnace below 50OF and
solution heat treated and quenched as follows:
hours at 940+IOF
2 hours at 950+1OF
16 hours at 980+1OF
The plates were then quenched in room temperature water
and aged for 8 hours at 360+5F.
Foundry C
The rigging for the plates produced by Foundry C is shown inFigure 4. Two 2.5-inch by 2.5-inch by 8-inch cast iron chills
placed end on end were used in the cope, and one 2-inch by 4-inch
18
Figure 4. Rigging System Used by Foundry C toProduce B201 Plates.
19
by 16-inch-long chill was used in the drag to obtain directional
solidification. Ten 1.5-inch-diameter insulated risers spaced at
approximately 4-inch intervals were used to feed the casting
during solidification. The mold cavity was gravity fed through
the gating system, which contained ceramic foam filters to trap
foreign material.
The plates were loaded into a furnace below 50OF and
solution heat treated as follows:
2 hours at 920+1OF
2 hours at 950+1OF
2 hours at 970+1OF
16 hours at 980+1OF
The plates were quenched into room temperature water and
then naturally aged at room temperature for 24 hours, followed by
artificial aging at 360+5F for 8 hours.
4.2 TEST PROCEDURES
The 1.25-inch-thick plates used for the verification
evaluations were comprehensively evaluated using the tests shown
in Table 3, and were radiographically inspected to assure all
materials were Grade B or better. The DADT test specimens were
excised randomly from the 1.25-inch-thick plates to determine the
between-plate and foundry-to-foundry mechanical property
variability.
For the 0.75-inch-thick verification plates, only tensile
and notched tensile tests were conducted. The results were
compared with the screening test data to assure that the
properties obtained from the optimized screening evaluation
plates [1] could be reproduced in the verification plates.
20
TABLE 3. VERIFICATION TESTS FOR D357-T6 AND B201-T7.
Number of Tests
D357-T6 B201-T7Test Specification Water Glycol Water
Quench Quench Quench
Tensile ASTM B557 30 10 30
Notched Tensile ASTM E602 30 10 30
Fracture Toughness ASTM B646(I) 6 3 6
Constant Amplitude ASTM E647 6 3 6Fatigue Crack Growth
Stress-Life Fatigue ASTM E466 12 6 12
Stress Corrosion ASTM G47 0 0 3
Strain-Life Fatigue ASTM E606( 2 ) 12 0 12and E466
Metallography ASTM E112 15 6 15
Chemical Analysis NA 15 6 15
(1)In addition to the ASTM 646 standard test, short bar testingwas performed using the method currently under review by ASTM.
(2)A new specification is under review by ASTM.
21
The composition of each plate was determined using
Inductively Coupled Plasma (ICP) analysis to confirm the foundry
melt results. The DAS of the D357 plates was measured using a
line intercept method [4] and the silicon particle morphology
was evaluated using an Omnicon 3500 image analyzer. Specimens
were excised from random locations throughout the plates.
4.3 EQUIVALENT INITIAL FLAW SIZE ANALYSIS
Durability and damage tolerance analysis requirements for
aircraft structure are defined in MIL-A-83444 and MIL-A-87221.
For damage tolerance analysis of wrought alloys, the presence of
a flaw is assumed so that defects that may occur in critical
areas of a part during manufacturing are accounted for. Castings
contain inherent flaws such as microporosity, microshrinkage, or
foreign material that may not be detectable by normal NDI
methods. A limited amount of these defects is acceptable in
premium quality aluminum castings for aerospace applications. An
equivalent initial flaw size (EIFS) must be determined for
castings to account for these inherent flaws. The EIFS of D357
and B201 was determined from the Task 2 (verification) fatigue
data using a method developed previously under a Northrop IR&D
Program [5]. The methodology for determining the EIFS and how it
may be applied to both durability and damage tolerance analyses
are summarized below. The damage tolerance analysis is included
in Section 5 of this report. The durability analysis is
scheduled to be concluded later in the program.
4.3.1 Determination of Equivalent Initial Flaw Size
Stress-life fatigue results for smooth round-bar specimens
(Kt = 1.0) were used to obtain an estimate of the inherent
material EIFS. The value of EIFS (ao), which initiates a crack
in a smooth round bar fatigue specimen, is based on the following
relationship [5].
22
(1-m/2) /m }1/ (-m12)
where ao = ai -(1-m/2)C(8AS$R/Q)N i Eq. 1
ai = a selected flaw size defining initiation
which is close to, but in excess of ao by
at least 0.01 inch.
m & C = the slope and intercept, respectively, of
the Paris law equation representing the
da/dn vs AK plot. For this analysis, the
da/dn test must be run at the same stress
ratio as the stress-life fatigue tests.
B = the stress intensity correction factor for
a semicircular surface flaw of a depth
equal to the average of ao and ai .
S= the elastic stress range of the fatigue
specimen.
Ni = Number of cycles for the fatigue crack to
grow to ai from ao .
Q = a crack shape factor, equal to
2.464-0.212(Smax/Oy)2 .
where Smax = the maximum fatigue stress.
Uy = yield strength.
The value of Ni is determined from the number of cycles to
failure using the following relationship:
m/2
(ai+ ao ) Aa(Nf/Ni) = 1+ m/2 Eq. 2(ai- aO) [ Inm ( ) /
a=ai \tj i I
where: a = the crack length at the midpoint of the
assumed interval, Aa.
23
B= the stress intensity correction factor at
crack length W.
The EIFS solution involves an interpolation between
Equations 1 and 2 because each one contains ao . However, an
adequate convergence occurs within seven or eight iterations
depending on the initial value of ai assumed.
4.3.2 Application to Durability Analysis
The durability analysis will be performed using the Northrop
sequence crack initiation program, LOOPIN8 [6] using data
generated from round bar strain-life fatigue test specimens.
These specimens have the same general geometry as those used to
obtain the EIFS (Section 4.3.1).
Using the average EIFS obtained as described in Section
4.3.1, the ratio between Ni and Nf for each alloy can be derived
from Equation 2 for a selected value of initiation flaw size
(ai). Hence, the failure lives obtained during the strain-life
tests can be converted to the cycles required to initiate a crack
of the selected size. Using these converted data, the LOOPIN8
program is then used to predict the life to initiate a crack from
the material EIFS (ao) to the predefined initiation flaw size
(ai)•
4.3.3 Application to Damage Tolerance Analysis
Damage tolerance analysis is based on the assumption that
crack growth starts from a "rogue" flaw. For wrought alloys, the
size of this flaw is generally defined [7] as a 0.050 inch corner
flaw at the edge of a hole. Reductions in the assumed value of
this flaw are negotiable between the supplier and the customer
based on a demonstration of manufacturing quality. However, if
the basic quality of the material is such that similar or greater
rogue flaws can exist regardless of the care taken to produce the
24
finished article, then the upper bound of the inherent material
equivalent initial flaw size becomes the limiting design factor.
The upper bound for the castings was derived from the EIFS
distribution obtained from smooth round bar fatigue test results
using a "sudden-death" analysis, as described below.
Sudden-Death Analysis
Sudden-death testing [8] is a method that can be used for
determining the upper bound defect size for the general
population based on the fatigue life results for specimens that
contain random distributions of casting defects. Crack
initiation will usually occur at the largest defect located at or
near the surface of the gauge section of the specimen. The
failure of the specimen can be assumed to be the "sudden-death"
failure of the set of defects contained within the surface volume
of the specimen. The largest defect is represented by the
shortest fatigue life.
The smooth round bar fatigue specimens used in the DADTAC
program had a 1.6-inch-long by 0.50-inch-diameter gauge section.
The volume of material in the critical initiation zone of these
specimens is equivalent to that obtained in a similar depth of
material at the edge of approximately 70 0.25-inch-diameter holes
drilled through a 0.25-inch-thickness. This estimate is based on
an assumption that the critical high stress concentration region
for a hole extends twenty degrees either side of the hole axis
aligned normal to the primary load direction. This is considered
to be a conservative assumption as a more localized stress
concentration would tend to increase the number of holes
represented by the smooth bar specimen.
According to the theory of ranking [8], the lowest value in
a set of 70 specimens clusters about the 1% probability level of
the general population. If the lowest life represents the
25
highest equivalent initial flaw size in each set of 70, then the
EIFS values obtained from an analysis of the round bar tests will
be clustered about the 99% probability level of the EIFS values
obtained for a typical defect population.
The ranked results are analyzed assuming a Weibull
distribution to obtain the median and upper 95% confidence limit
statistics of the "sudden death" population. The median and
upper 95% confidence limit of the general population are obtained
by translating the "sudden death" median line and confidence
limit upward until the Weibull median value of the "sudden death"
line coincides with the 99% probability value.
EIFS values for castings obtained from different vendors
were grouped using variance analysis [9). An estimate of the
maximum inherent material rogue flaw size was obtained from the
99.9% probability and 95% confidence value read from the above
derived general population distribution. If this value exceeds
the current requirements of MIL-A-83444 it indicates that the
basic material quality is more critical for the castings
evaluated under the DADTAC program than the upper flaw sizes
currently assumed for wrought alloys based on manufacturing
quality. A value less than the current MIL-A-83444 specification
requirement would indicate that castings can be considered for
use in a damage tolerant design by applying the same criteria
used for wrought alloys.
26
SECTION 5
RESULTS AND DISCUSSION
The overall objective of the verification subtask was to
determine the DADT properties of D357-T6 and B201-T7 made to the
specifications developed in the screening subtask. The results
for the two alloys are described in the following subsections.
5.1 D357-T6 RESULTS
The main emphasis for D357 was to evaluate cast plates that
were quenched in room temperature water. However, a few D357
plates were quenched in glycol (GQ) to determine if the quench
medium significantly influenced the mechanical properties. All
other process parameters for the GQ plates were the same as for
the water-quenched (WQ) plates.
For the WQ plates, only those that fully met the specified
microstructural and tensile properties as well as radiographic
quality and chemical composition, were used for evaluating DADT
properties. To fully evaluate the effect of the quench, all of
the GQ plate test results were compared with data from the WQ
plates. However, only data from those GQ plates which fully met
all the specification requirements were included in the subse-
quent DADT analyses (e.g. EIFS).
All the WQ and GQ D357-T6 plates from Foundry A met the S
specification requirements. The DADT test specimens, therefore,
were excised from these plates at random locations. For Foundry
B, two WQ plates did not meet the tensile property requirements;
both had one low elongation value (<3%) out of the two tensile
specimens tested from each plate. Therefore, only the remaining
three plates were used for the DADT characterization described in
this section. The evaluations for Foundry C material were not
complete at the time this report was prepared.
27
5.1.1 Composition
A combined total of sixteen plates (1.25-inch- and 0.75-
inch-thick) cast from ten different melts by two foundries were
used for the verification evaluations. The composition of each
of these melts and the specification ranges for each element are
listed in Table 4. The compositions were within the range speci-
fied from the screening subtask of this program (Table 1). The
melt analyses were confirmed by ICP chemical analyses on samples
taken from the plates.
5.1.2 Microstructure
The microstructure of the D357-T6 plates was characterized
using optical, transmission and scanning transmission electron
microscopy (TEM and STEM), and scanning electron microscopy
(SEM). The DAS and silicon particle morphology of each plate
were determined. A typical optical micrograph of D357-T6 is
shown in Figure 5.
The DAS of specimens taken from various locations throughout
the plates ranged from 0.0008 inch to 0.0020 inch, which is
within the specification requirement (<0.0024 inch).
The average DAS values, as well as the Si particle area,
spacing, and aspect ratio for WQ and GQ material from both
foundries are listed in Table 5, along with a comparison to the
corresponding data obtained previously for the screening
material. The DAS for material from the two foundries was
similar, with all values being below the specification maximum of
0.0024 inch. The average DAS of the verification material was
slightly less than that of the screening material. The average
DAS for the GQ plates from both foundries was less than that for
the WQ material. The GQ and WQ plates were made to the same
specification using identical processing conditions except for
the quench medium. There was no variation in chill or mold
materials which would affect the DAS. Therefore, the consistent
28
A
TABLE 4. D357 VERIFICATION MATERIAL MELT COMPOSITIONS.
COMPOSITION (wt %)
Fe Si Mg Ti Mn Be Na Sr Al
0.03 7.03 0.6 0.12 0.00 0.050 0.002 - Bal.0.03 6.99 0.6 0.11 0.00 0.070 0.003 - Bal.0.03 7.27 0.6 0.10 0.00 0.060 0.009 - Bal.0.04 7.37 0.6 0.11 0.00 0.050 0.012 - Bal.0.11 6.67 0.6 0.16 0.01 0.061 - 0.008 Bal.0.10 7.35 0.6 0.15 0.01 0.070 - 0.014 Bal.0.11 6.80 0.6 0.17 0.01 0.059 - 0.011 Bal.0.10 7.07 0.6 0.15 0.01 0.069 - 0.014 Bal.0.11 7.20 0.6 0.16 0.01 0.070 - 0.007 Bal.0.10 7.07 0.6 0.15 0.01 0.069 - 0.014 Bal.
AMS 4241 Specification Range
<0.12 6.5- 0.55- 0.10- <0.1 0.04- * * Bal.7.5 0.6 0.20 0.07
• Not specified in AMS 4241; however, those amounts used arewithin the 0.05 wt. % limit set for "other" elements.
29
q
V I- 4'
V '
- A4-
/hi ,.
A'
S50X As polished
Figure 5. Typical D357-T6 Microstructure.
S1
30
TABLE 5. D357 DAS AND Si PARTICLE DATA.
Si Particle Morphology
Foundry Quench DAS Area Spacing Aspect % Porosity(inch) (Mm2 ) (Am) Ratio
A Water 0.0012 16 38 1.6 0Glycol 0.-,,,O 12 36 1.6 0
B Water ,.0018 18 44 1.6 0.083Glyco-. 0.0013 23 40 1.6 0
Average Water 0.0015 14 41 1.6 0.042Glycol 0.0011 18 38 1.6 0
Average of 0.0016 20 36 1.8 0.023ScreeningData [1]
31
difference in the DAS for the GQ and WQ plates is considered to
be coincidental and not a result of processing differences.
The silicon particle data for the verification material from
the two foundries were also similar, and were similar to those
obtained previously for the screening material. The percent
porosity, which was determined using the optical metallography
specimens, was low for all of the verification plates, as it was
in the screening plates. An additional indication of the amount
of gas porosity was obtained by measuring the residual hydrogen
present in samples of the cast material. The average hydrogen
content of samples from both foundries was 0.65 ppm.
The microstructure of randomly selected samples was
evaluated using SEM, TEM, STEM, SAD patterns, and EDX analyses.
Four different phases were observed, although the exact
stoichiometry of three could not be determined.
A typical SEM micrograph of D357-T6, Figure 6, shows three
phase morphologies in addition to the aluminum matrix. Using EDX
analyses, the shapes identified by Numbers 1 and 2 were
determined to be Al-Fe, but the stoichiometry could not be
determined. The phase identified by Number 3 was determined by
EDXA to be Si. Smaller Si particles were also observed in the
TEM micrograph, Figure 7. An X-ray map of the microstructure
confirmed that these particles were Si. Long needle-shaped
particles were also observed in the microstructure of Figure 8,
and were shown by X-ray mapping to contain Ti. Mg and Fe X-ray
maps of this microstructure were also generated, but neither
element was detected. A closer examination of the Ti-containing
needles was conducted to determine their stoichiometry. Figures
9a and b show a typical needle and its SAD pattern, respectively.
Examination of the pattern, as well as two others obtained at
different orientations, did not reveal the composition or
stoichiometry of the phase. Another phase is also seen in
32
Figure 6. Backscatter SEM Figure 7. TEN Micrograph of
Micrograph of D357-T6. D357-T6 Showing Si Particles.
100
Figure 8. TEN Micrographof D357-T6 Showing a Needle-Shaped Phase Containing Ti.
S 33
nS
Figure 9a. TEM Micrograph Figure 9b. Selected Areaof D357-T6 Showing a Needle- Diffraction Pattern of theShaped Particle Containing Ti Needle Phase of Figure 9a.and Smaller Mg2Si Particles.
34
Figure 9a, which was identified by EDXA to be the main
strengthening phase Mg2Si.
Sodium was used to modify the as-cast microstructure of the
specimen that was examined in the TEM. However, X-ray mapping
and diffraction patterns of various areas of the microstructure
failed to locate Na. Some controversy exists about the exact
modification mechanism caused by Na. However, one theory [20] is
that modification is accomplished through an interaction between
Na and P, an impurity element. In the absence of Na, P forms
AlP, which promotes the formation of large Si particles. When Na
is added, the compound NaP is formed, which suppresses this
tendency, resulting in a finer microstructure.
5.1.3 Tensile Properties
The tensile properties of D357-T6 plates from Foundries A
and B are shown in Table 6. Tests were conducted on both water-
and glycol-quenched material from the 1.25-inch-thick plates. A
limited number of tensile tests were also conducted on the 0.75-
inch-thick plates (water-quenched only). These smaller plates
were evaluated to assure that those properties obtained for the
optimized material during the screening evaluations could be
reproduced in plates of the same size produced to the AMS 4241
specification. The combined average tensile properties for the
0.75-inch-thick verification plates from Foundries A and B were
53.6 ksi ultimate strength, 45.3 ksi yield strength, and 5.9
percent elongation. These values were similar to the results
from the screening tests, i.e. 51.4 ksi, 43.4 ksi, and 5.1,
respectively.
The average and minimum tensile properties for the 1.25-
inch-thick water-quenched material from Foundry A (54/45/5.9)
were slightly higher than those from Foundry B (52/45/4.5). All
of the individual test specimens from the Foundry A plates met
the 50/40/3 tensile property requirements. Two Foundry B plates
351
TABLE 6. TENSILE PROPERTIES OF THE 1.25-INCH-THICK, WATER-
AND GLYCOL-QUENCHED* D357-T6 PLATES.
Quench** UTS (ksi) YS (ksi) El (%)Foundry Medium Avg. Range Avg. Range Avg. Range
A Water 54 51-55 45 44-47 5.9 3.5-8.5Glycol 53 51-55 43 42-44 6.6 5.5-9.0
B Water 52 50-53 45 44-46 4.5 3.0-5.6Glycol 49 48-50 42 41-43 3.4 2.5-4.3
Average Water 53 45 5.2
Glycol 51 43 5.0
Target Min. 50 40 3.0
*25% Glycol Solution**Room Temperature
36
36
each had one tensile test result that had a percent elongation
less than the 3.0% minimum requirement. Subsequent DADT
evaluations of material from Foundry B were conducted only on
those plates that fully met the tensile property specification
requirements. The tensile property data for Foundry B material
shown in Table 6 include only those WQ plates that were within
the specification.
The average and minimum tensile property data for GQ plates
from both foundries are also shown in Table 6. Strength levels
are in general slightly lower than those for the WQ material.
The reduction in strength is probably associated with a lower
level of Mg remaining in solution following the quench, resulting
in less Mg2Si, the strengthening phase, following artificialaging. Two of 12 elongation values for the GQ plates were below
the 3.0 percent minimum.
5.1.4 FatiQue Properties
5.1.4.1 Smooth Stress-Life
The stress-life data for D357-T6 (kt=l.0) from Foundries A
and B are shown in Figure 10 along with a comparison to 7075-T73
£11], a wrought alloy commonly used in fatigue critical aircraft
applications. The fatigue lives of material from Foundry A are
longer than those for Foundry B for a given stress level. The
fatigue lives of the GQ and WQ material are essentially the same.
The average fatigue life of D357-T6 is less than that of 7075-T73, and is presumably due to lower D357-T6 strength and the
larger size of the inherent defects found in castings, which will
tend to reduce the initiation life, as discussed in Section
5.1.4.4.
5.1.4.2 Smooth Strain-Life
The strain-life data are shown in Figure 11 as log strain
amplitude vs. log of the number of cycles to failure (Nf).
37
70
R-0.1 ; Kt- 1.0
*6705-7 FOUNDRY W A'R GLYCOL60B
A A
50
O0 0 A
w 4 0 0 ft
U,I-A
030 ON A AE
• 20- •-E10
10
0 I l I II III I I I I It I I I II III10 4
05
06
10710 ~ 10 ~ 10
Log N f
Figure 10. D357-T6 Stress-Life Fatigue Data.
10.2
R= -1.0; Kt= 1.0
0 Foundry Symbol
A AB 0
0.)
El A
CL
ECc A
A
0 0 A-J
1 0 3 lJ
10 1 102 10 3 10 4 10 5 10 6 10 7
Log N f
Figure 11. D357-T6 Strain-Life Fatigue Data.
38
The specimens were tested using an R ratio of -1.0, with strain
amplitudes ranging from 0.010 to 0.002, which resulted in fatigue
lives ranging from approximately 50 to 1,000,000 cycles. These
data were generated for use in the durability analysis later in
the program (described in Section 4.3.2).
5.1.4.3 FatiQue Crack Growth Rate
The range of constant amplitude fatigue crack growth rate
(FCGR) data for WQ and GQ material from Foundries A and B are
shown in Figure 12. Also included are data for A357-T6 obtained
under a previous Northrop program [12], and 7075-T7351 plate
[13]. The DADTAC D357-T6 data are similar to the previous A357-
T6 data and are slightly better than that of 7075-T7351. Based
on these data, the difference in the total fatigue life between
the cast and wrought materials (Figure 10) can be attributed to
the ease of crack initiation. Fatigue cracks probably initiate
more readily in castings than in wrought material, due to the
presence of inherent defects such as dross and/or porosity. Each
fatigue life specimen was fractographically examined to determine
the crack initiation site. Out of 12 specimens examined, 11
initiated a fatigue crack from a defect located at or near the
surface. No clearly defined defect was associated with the
fatigue crack initiation of the twelfth specimen. The results
are shown in Table 7. All of the defects were porosity and/or
dross. Typical sites are shown in Figures 13a and b.
5.1.5 Fracture Toughness
The fracture toughness data for WQ and GQ D357-T6 are shown
in Table 8. Compact tension, short bar, and NTS/YS tests were
conducted. A short bar test specimen was removed from each of
the tested compact tension specimens to obtain a direct
correlation between the two test methods. Invalid KQ results
39
1o-4
Northrop IR&D (12) 0.
7075-T7351/* Plate [13)
RegressionQ Fit of Region 11o da/dN Data
m-4.988 I10r c.2.46x10-11 J /
Range of DADTAC 0357-T6 Data
10.10 100AK (kSri(7i)
Figure 12. D357-T6 Fatigue Crack Growth Rate Data(R=0.l; lab air).
TABLE 7. FATIGUE CRACK INITIATION SITES FORD357-T6 FATIGUE SPECIMENS.
Number of Fatigrue Specimens With Crack Initiation Site
Dross Pore Both Dross and Pore Other*
4 4 31
*No Observable Defect
40
a. SEM M~icrograph of a Pore.
~~~~~ JA t- ~ 'n'-s
ii~i A :b.~~~~~~~~ Backsatte S irgap fFrig aeil
b.ur 1a3.ate Ty EMa Crcrogritahio Foresignaeil
D357-T6 Fatigue Specimens.
41
TABLE 8. AVERAGE D357-T6 FRACTURE TOUGHNESS DATA.
Quench Fracture Toug~hness (ksiN in.) NTS YSMedium KIC KQ KIV* (ksi) (ksi) NTS/YS
Water - 24 24 62 45 1.38
Glycol** 22 -22 57 42 1.36
*Short Bar Test**25% Glycol Solution
42
were obtained from the compact tension tests for all the WQ
material due to excessive crack front curvature, which is often
observed in castings due to residual stresses incurred during
quenching. Valid Kic results were obtained for GQ plates,
presumably because of the slower cooling rate associated with the
glycol.
For both foundries, an average KQ of 24 ksiV'Th was obtained
for WQ material by both compact tension and short bar (KIV) test
methods. This fracture toughness value corresponded to a NTS/YS
ratio of 1.38. The fracture toughness of the GQ material was
slightly less than that of the WQ material, 22 ksi in for both
compact tension (KIc) and short bar (KIV) test methods. The
average NTS/YS ratio of the GQ material was slightly less than
that of the WQ material, 1.36 vs. 1.38. Similar to the WQ
material, both foundries had the same average fracture toughness
values for the GQ plates.
The ranges of the compact tension and short bar fracture
toughness values and their correlations with yield strength are
evident by comparing all the test data shown in Figure 14. The
fracture toughness ranged from approximately 21 to 26 ksi in,
with the lowest values (21 and 22 ksi vin, KIC) being obtained
for the GQ material. There was no clear correlation between the
fracture toughness and yield strength, probably because the yield
strength range was relatively narrow. Other work [14] has shown
that there is an inverse relationship between fracture toughness
and yield strength.
5.1.6 Eguivalent Initial Flaw Size Analysis
The method for determining the equivalent initial flaw size
was described in Section 4.3.1. The calculated values for
individual specimens excised from material from the two foundries
are shown in Table 9, along with the corresponding fatigue life
and maximum applied stress. Testing of specimens that did not
fail after 5x10 6 cycles was terminated and an EIFS was not
0 43
30
Glycol Water K oQuenched Quenched - Kic
c 25IM
4 45
4 44
--
u
20
40 45 50Yield Strength (ksl)
NOTE: All K10 values for glycol - quenched material.KN obtained from short bar test.
Figure 14. Relationship Between Fracture Toughness andYield Strength for D357-T6.
44
TABLE 9. D357-T6 EQUIVALENT INITIAL FLAW SIZE DATA.
Maximum Cycles toFoundry Stress Failure EIFS
(ksi) (x103 ) (inch)
A 30 4069 0.002730* 1474 0.005235 220 0.010345 50 0.011140* 50 0.015940 50 0.016220* 5300+ -
Avg. 0.0102
B 30* 254 0.015040* 45 0.016730 213 0.016820* 982 0.022345 14 0.0228
25 223 0.0267
Avg. 0.0201
*Glycol-quenched+No failure
45
calculated.
The data indicate that the EIFS for Foundry A is less than
that for Foundry B. This is consistent with the S/N fatigue data
shown in Figure 10; the results for test specimens from Foundry A
material are at the higher end of the overall band of data. The
EIFS data were then analyzed as described in Section 4.3.3 to
determine the value of the maximum inherent material rogue flaw
size (99.9% probability and 95% confidence) that should be used
to design a cast DADT critical component to a specific life. The
data are presented as a Weibull distribution in Figure 15. For
D357-T6, the maximum flaw size was determined to be 0.030 inch,
which is less than the 0.050 inch flaw that is typically assumed
for damage tolerance analysis of wrought materials [7]. A defect
of this size may be detected in material up to about 3-inches-
thick using X-ray radiography at a 1% sensitivity. Based on this
information, Grade B or better D357-T6 castings up to the maximum
thickness evaluated under the DADTAC program (1.25 inch) can be
considered for use in a damage tolerance design.
The fracture surfaces of the fatigue coupons were examined
to determine the size of the crack initiation sites for a
comparison to those derived from the EIFS analysis, Table 10.
The measured values were obtained by determining the semicircular
area within which the flaw would fit with the center of the
semicircle on the surface of the specimen. Measured values were
not obtained for all twelve fatigue specimens because exact
identification of the complete crack site was not possible due to
the presence of multiple defects. The derived values are within
an order of magnitude of the measured values. Discrepancies
between the measured and derived values are most probably due to
the inability to accurately account for the effects of the shape
(sometimes very complex) and orientation of the crack initiation
sites relative to the specimen surface. The difficulty of
accurately measuring defects is illustrated in Figures 16a and b.
The measured defect size is indicated by the semicircle enclosing
46
99.9
99 C989590
80 1 A V *70 4" V
60 -- /
50'1 A /^ 9%Cn pce
40 _4
30
20
510C
4
3
0.110 100
Equivalent Initial Flaw Size (0.001 inch)
Note: Weibull mean of the sudden death data (A)clusters about the 99% probability level of thegeneral population (B). The maximum (99.9%probability with 95% confidence) expected flawis indicated by point (C).
Figure 15. Weibull Distribution of D357-T6 EIFSData.
47
TABLE 10. COMPARISON OF PREDICTED AND MEASUREDEQUIVALENT INITIAL FLAW SIZE FOR D357-T6.
EIFS (in)Specimen
Predicted Measured
Vill 0.0027 0.0175V115 0.0052 0.0375VII6 0.0159 0.0343V209 0.0229 0.0096V210 0.0267 0.0162V211 0.0150 0.0062V213 0.0223 0.0137
48
Dross
75XSpecimten Surface
a. Dross.
9 A EI
75X Specimen Surface
b. Pore.
S Figure 16. Measurement of Defect Size in D357-T6.
* 49
each defect with the center of the semicircle placed on the edge
of the specimen. (A semicircle was used because DADT analyses
assumes the presence of a semicircular flaw in the material.) It
can be clearly seen that the semicircle encloses much more
material than the defect, and that the size of the semicircle is
highly dependent upon the aspect ratio of the defect. In
contrast, the EIFS analysis determines the semicircular area
which has the same effect on fatigue properties as a highly
aspected defect. Highly aspected defects can cause the measured
defect size to be either greater than or less than the predicted
defect size (the EIFS). From this comparison, it was concluded
that the approach being used to predict the EIFS from smooth
fatigue data is viable.
5.2 B201-T7 RESULTS
5.2.1 Composition
Twenty plates cast (both 1.25-inch-thick and 0.75-inch-
thick) from eight different melts by three foundries were used
for the B201 verification evaluations. The composition of each
of these melts and the ranges for each element are listed in
Table 11. All of the compositions were within the specified
range (AMS 4242). Reported compositions of the plates were
confirmed by ICP tests at Northrop.
05.2.2 Microstructure
The microstructure cf the B201-T7 plates was characterized
using optical, TEM, STEM, and SEM methods, and the grain size and
percent porosity of each plate were determined.
The grain size at random locations in plates from all three
foundries ranged from 0.0024 inch to 0.0039 inch. A typical
micrograph of the B201-T7 microstructure is shown in Figure 17.
50
TABLE 11. B201 VERIFICATION MATERIAL MELT COMPOSITIONS.
COMPOSITION (wt %
Cu. Ag Mg Mn Ti Fe Si Al
4.59 0.56 0.27 0.28 0.19 0.02 0.04 Bal
4.59 0.56 0.27 0.28 0.19 0.02 0.04 Bal.4.55 0.55 0.30 0.28 0.18 0.01 0.02 Bal.4.75 0.40 0.27 0.22 0.17 0.01 0.05 Bal.4.75 0.44 0.24 0.33 0.16 0.02 0.05 Bal.4.93 0.64 0.28 0.28 0.28 0.04 0.04 Bal.4.9 0.69 0.28 0.28 0.28 0.04 0.04 Bal.
4.63 0.57 0.25 0.31 0.18 0.02 0.01 Bal.
AMS 4242 Specification
4.5- 0.40- 0.20- 0.20- 0.15- <0.05 <0.05 Bal.5.0 0.80 0.30 0.50 0.35
AS
50X" Kelr etc
Figur 17 Tyia K07Mcosrcue/5
The microstructure was examined further by SEM, STEM, and
TEM. SAD diffraction patterns and EDXA of various phases were
taken, leading to the identification of five different phases.
The most prominent of these five was the strengthening phase
CuA12 , (theta), Figure 18, found as small platelets on the (100)
plane of the aluminum matrix. A grain boundary phase and
associated precipitate-free zone were also observed in Figure 18.
EDX analysis of the grain boundary area identified the presence
of Al and Cu, suggesting the phase was equilibrium CUAl2. CuAl2was also seen as a large blocky constituent phase, Number 2 in
Figure 19. Other large constituents were also observed in
Figure 19. The constituent labeled Number 1 contained Cu and Fe.
The constituents indicated by Numbers 3 and 4 were rich in Ti,
and those indicated by Numbers 5 and 6 contained Cu, Fe, and Mn.
Two grain boundary phases were also observed in addition to
the equilibrium CuAI 2 phase. One was the Al, Cu, Fe, and Mn rich
phase mentioned previously, Number 1 in Figure 20; the other was
rich in Al, Cu, and Mn, Number 2. A large blocky Al, Cu, and Mn
phase was also observed in the matrix, Figure 21. Figure 22
shows another Al-Ti phase, which was either AI3Ti, A124Ti6 , or
Al23Ti9.
The CuAI2 phase identified above has been reported [15] for
alloys containing Ag to be omega, an altered form of theta, the
main strengthening precipitate in Al-Cu alloys. The addition of
Ag to Al-Cu alloys promotes the formation of the omega
strengthening precipitate during aging above about 210F and
causes a marked increase in age hardening. Silver was also shown
by regression analyses [1] to increase the strength of B201-T7.
Omega essentially replaces theta and is thought to be a
monoclinic or hexagonal form of the theta phase, which itself is
body-centered tetragonal. Omega forms on the (111) plane of the
matrix as a uniform dispersion of large, very thin, hexagonally
shaped plates, and is thought to be coherent with the matrix
along the (111) plane and at the ends of the precipitate. Omega
52
Figure 18. TEN Micrograph of B201-T7 Showing theTheta Phase.
I 530
Figure 19. Backscatter Figure 20. STEM Micrograph ofSEM Image of B201-T7. B201-T7 Showing a Phase Rich in
Al, Cu, Fe, and Mn.
Figure 21. TEM Micrograph Figure 22. TEM Micrographof B201-T7 Showing a Phase of B201-T7 Showing a PhaseRich in Al, Cu, and Mn. Rich in Al and Ti.
54
may be more stable than theta, as evidenced by creep testing
[15). In addition, the precipitation mechanism associated with
the omega phase seems to be different from that of the theta
phase.
Scanning diffractometry of B201 was also conducted to
identify phases. Analysis of the diffraction pattern led to the
identification of only two phases, A1203 and delta (Al4Cu9).
Phases containing other alloying elements were not detected due
to the relatively small amount (<l wt percent) of all alloying
elements in B201 except for Cu.
5.2.3 Tensile Properties
The average and minimum (indicated by the ranges) tensileproperties of the 1.25-inch-thick B201-T7 plates from all three
foundries are shown in Table 12. Test results were also obtained
for the 0.75-inch-thick plates to provide a direct comparisonwith those obtained in the screening portion of Task 2. Average
tensile properties of 66.5 ksi ultimate strength, 59.1 ksi yieldstrength, and 8.3 percent elongation to failure were obtained for
the 0.75-inch-thick verification plates with minimum values of
64/55/6.5. These properties are similar to those obtained for
the screening material (65/57/7).
All the individual tensile properties for the 1.25-inch-thick verification plates from all three foundries exceeded the
specified minimum of 60 ksi ultimate strength, 50 ksi yield
strength, and 3 percent elongation. Material from Foundry B hadthe highest strength; material from Foundries A and C had similar
properties. Stress corrosion tests were conducted (Section
5.3.6) and confirmed that the alloys were in the T7 condition.
55
TABLE 12. TENSILE PROPERTIES OF 1.25-INCH-THICKB201-T7 PLATES.
UTS (ksi) YS (ksi) El (%)Foundry Avg. Range Avg. Range Avg. Range
A 66 65-67 57 54-58 8.7 8.3-11.0
B 70 68-73 63 60-66 8.2 7.5-8.8
C 67 64-70 60 58-62 8.3 6.5-9.5
Average 68 60 8.4
Target Min. 60 50 3.0
56
5.2.4 Fatigue Properties
5.2.4.1 Smooth Stress-Life
The stress-life (kt=l.0) data for B201-T7 from the three
foundries are shown in Figure 23, along with a comparison to data
for A201-T7 material obtained under a previous Northrop/Navy
contract [16], and 7075-T73 [11], an alloy commonly used in
fatigue-critical aircraft applications. The fatigue lives for
B201-T7 from the three foundries are similar and are essentially
the same as the previous results [16]. However, the fatigue life
for a given stress was shorter than that of 7075-T73.
5.2.4.2 Smooth Strain-Life
The strain-life data, plotted as log strain amplitude vs.
log of the number of cycles to failure (Nf), are shown in Figure
24. The specimens were tested using an R ratio of -1.0, with
strain amplitudes ranging from 0.008 to 0.002. These strain
amplitudes resulted in fatigue lives ranging from approximately
500 to 200,000 cycles. These data were generated for use in the
durability analysis to be conducted later in the program
(described in Section 4.3.2).
5.2.4.3 Fatigue Crack Growth Rate
The range of fatigue crack growth rate (constant amplitude)
data for B201-T7 from the three foundries is shown in Figure 25.
Also included are data for A201-T7 tested under a Northrop/Navy
contract [16], and 7075-T7351 plate [13], an alloy used in fa-
tigue critical aircraft applications. The DADTAC program data
are similar to the Northrop/Navy contract A201-T7 data. However,
B201-T7 had a slightly slower fatigue crack growth rate than
7075-T7351. Because the FCGR for castings and 7075-T7351 is
similar, the difference in total fatigue life (Figure 23) can be
attributed to the reduced crack initiation resistance of the cast
57
70 A-0.1 ;K, . 1.0
60
50
40 0 a
(i))
30 -C /- - -M - - -- ---------Ea
20 - Northrp- Navy Contract 1161 0 0 Zt
0 b*
10
104 105 LogNf 106 107
Figure 23. B201-T7 Stress-Life Fatigue Data.
10-2
R= -1.0; Kt= 1.00 Foundry Symbol
A AB C3
C 0
-~ 0 £1
CL A 0'
Ec 00 A
0) 00
1 3 1 L L 16
101 102 103 104 105 106
Log Nf
Figure 24. B201-T7 Strain-Life Fatigue Data.
58
IOA
7075-T7351 Plate (13)
10 Not hrop-Navy/Contract [161
o aI
RerIso Range of DADTAC B20 1-Ti Data
Fit of Region 11da/dN Datam-3.882c-3.74010'1
0
110 100
AK (kSiIlni)
Figure 25. B201-T7 Fatigue Crack Growth Rate Data
(R=0.l; lab air).
590
material, which is presumably due to the presence of inherent
defects such as dross and/or porosity.
Each fatigue life specimen was fractographically examined to
determine the crack initiation site(s). Out of eight specimens
examined, fatigue cracks in seven of them initiated at defects.
It was not possible to detect a defect associated with the crack
initiation in the remaining specimen. The defects were eithergas porosity, residual shrink porosity, or foreign material. A
typical foreign material defect is shown in Figure 26.
5.2.5 Fracture TouQhness
Average compact tension and short bar fracture toughnessvalues, as well as NTS/YS ratios, of B201-T7 are shown in Table
13. The short bar test specimens were excised from the brokencompact tension specimens to provide a direct comparison of the
two test methods. An average KQ value of 39.4 ksiJ in was
obtained from compact tension testing, which is similar to the
average KIV value of 40.4 ksi in obtained from the short bartest. The average KQ values for Foundries A, B, and C were 43,
29, and 47 ksi J-h, respectively.
Invalid KQ results were obtained for all the compact tensiontests due to excessive crack front curvature, similar to the WQ
D357-T6 specimens (Section 5.1.5).
Individual specimen data and the relationship between
fracture toughness and yield strength is shown in Figure 27. The
fracture toughness decreased with increasing yield strength andranged from 27 to 46 ksi in. The corresponding short bar
results ranged from 27 to 58 ksi vin.
60
4j 4
Figure 26. Foreign Material on a B201-T7Fatigue Specimen Fracture Surface.
61
TABLE 13. B201-T7 FRACTURE TOUGHNESS DATA.
FRACTURE TOUGHNESS (ksi Vi NTS YS NTS/YSKQ KIV* (ksi) (ksi)
39.7 40.4 87.3 60.3 1.45
*Short bar test
60
M~ K 1;
U) 45 a)MCC
0 40
5-
0
25.
55606Shr a es il trn t ki
Shor BarTes Yield Strength ofksi) T7
62
5.2.6 Stress Corrosion
Stress corrosion tests were conducted on B201 from all threefoundries. Direct tension, alternate immersion tests were
conducted; specimens were exposed to a 3.5% NaCl solution for 30
days at 75% of the material yield strength. No failures were
experienced after the 30 day exposure, indicating that each plate
had been heat treated to the T7 (overaged) condition.
5.2.7 Equivalent Initial Flaw Size Analysis
The method for determining the equivalent initial flaw size
was described in Section 4.3.1. The calculated values for
individual specimens excised from material from each of the three
foundries are shown in Table 14, along with the fatigue life and
maximum applied stress. Testing of specimens that did not fail
after 5x10 6 cycles was terminated and an EIFS was not calculated.
The data indicate that the EIFS for material from Foundry A
is less than that for Foundries B and C material. This is
consistent with the stress-life fatigue data shown in Figure 23;
the results for Foundry A material are at the higher end of the
overall data band. Data for Foundries B and C are nearly
equivalent.
The EIFS data were then analyzed as described in Section
4.3.3 to determine the value of the maximum inherent rogue flaw 0
size (99.9% probability and 95% confidence) that should be used
to design a cast DADT critical component to a specific life. The
data are presented as a Weibull distribution in Figure 28. For
B201-T7, the maximum flaw size was determined to be 0.042-inch, 0which is less than the 0.050 inch flaw that is typically assumed
for damage tolerance analysis [7]. A flaw of this size can be
detected in material up to about 4-inches-thick by X-ray
radiography using 1% sensitivity. Based on this information, 0
Grade B or better B201-T7 castings up to the maximum thickness
63
TABLE 14. B201-T7 EQUIVALENT INITIAL FLAW SIZE DATA.
Maximum Cycles to EIFSFoundry Stress Failure (inch)
(ksi) (x103 )
A 40 187 0.003745 309 0.001430 5,000* -20 5,000* -
Avg. 0.0025
B 25 39 0.011130 28 0.040435 188 0.006720 5,000* -
Avg. 0.0194
C 40 16 0.028920 118 0.008620 573 0.016030 168 0.011815 5,000* -
Avg. 0.0163
*No Failure
64
99 _ __
60 __ _
20 50 Con idence
to _
C)~IC),
2
Equvaen Iniia FlwSz00.0 nh
Note: Weibull mean of the sudden death data (A) clustersabout the 99% probability level of the general population(B). The maximum (99.9% probability with 95% confidence)expected flaw is indicated by point (C).
Figure 28. Weibull Distribution of B201-T7 EIFS Data.
65
evaluated under the DADTAC program (1.25 inch) can be considered
for use in a damage tolerance design.
The fracture surfaces of the fatigue coupons were examined
to determine the size of the crack initiation sites. A compari-
son of these measured values and those derived from the EIFS
analysis is shown in Table 15. The measured values were obtained
by determining the semicircular area within which the flaw would
fit, with the center of the semicircle on the surface of the
specimen. The derived values are within an order of magnitude cf
the measured values. Discrepancies between the measured and
derived values are probably due to the inability to accurately
account for the effects of the shape and orientation of the crack
initiation sites as was discussed previously in Section 5.1.6.
TABLE 15. COMPARISON OF PREDICTED AND MEASUREDEQUIVALENT INITIAL FLAW SIZE FOR B201-T7
EIFS (in)Specimen
Predicted Measured
V156 0.0037 0.0025V256 0.0111 0.0037V355 0.0118 0.0366V356 0.0086 0.0316V357 0.0160 0.0075
66
SECTION 6
CONCLUSIONS
1. Tensile properties of the verification material producedusing the specifications derived from the screening portion
of Task 2, which were similar to AMS specifications 4241 and
4242 except for the addition of a Si modifier for D357-T6,
were similar to those of the screening plates. Thus, the
specifications can be met by the foundries, and are valid for
material of a thickness up to that evaluated in this program
(1.25-inches-thick).
2. The fatigue crack growth rates of D357-T6 and B201-T7
castings are similar, but are slightly slower than those of
7075-T7351.
3. Based on an evaluation of the smooth fatigue (kt=l.0)
specimen fracture surfaces, the overall fatigue life, and the
FCGR, crack initiation occurs sooner in castings than in
7075-T7351 due to the presence of inherent defects.
4. The equivalent initial flaw size for D357-T6 (0.030-inch) and
B201-T7 (0.042-inch) is less than the value assumed for
wrought alloys (0.050-inch). For castings up to 1.25-inch-
thick, these alloys can be considered for damage tolerance
applications based on the existing damage tolerance
specifications (MIL-A-83444 and MIL-A-87221).
5. The average fracture toughness (KQ) of D357-T6 was adequate
(24 ksivin). Though the average value for B201-T7 was
excellent (40 ksiv/in), there was significant foundry-to-
foundry variability, ranging from 20 to 47 ksi in.
6. Excellent agreement between results of the compact tension
and short bar fracture toughness tests was attained.
67
SECTION 7
REFERENCES
1. M.W. Ozelton and G.R. Turk, "Durability and Damage Tolerance
of Aluminum Castings," First Interim Report, Contract No.
F33615-85-C-5015, September, 1987.
2. Aerospace Materials Specification, AMS 4241, 1986, Society
of Automotive Engineers, Warrendale, PA 15096.
3. Aerospace Material Specification, AMS 4242, 1986, Society of
Automotive Engineers, Warrendale, PA 15096.
4. "Determination of Dendrite Arm Spacing in Aluminum
Castings," Northrop Process Specification IT-72.
5. P.G. Porter, "A Unified Approach to Crack Initiation and
Growth Analysis," Northrop Report No. NOR 87-102, October,
1987.
6. "Northrop Fatigue/Damage Tolerance Manual," Vol. I, Section
III.
7. "Airplane Damage Tolerance Design Requirements," MIL-A-
83444, July, 1974.
8. Lipson, C. and Sheth, N.J., "Statistical Design and analysis
of Engineering Experiments," Section 5.5, McGraw-Hill, 1973.
9. "Fatigue-Endurance Data," Vol 2., Aeronautical Series of
Engineering Sciences Data, Item No. 68016, The Royal Aero-
nautical Society, London, England, 1968.
10. Polmear, I.J., "Light Alloys - Metallurgy of the Light
Metals," ASM International, Metals Park, OH, 1982, p. 119.
69
11. Northrop IR&D Program 85-R-1098.
12. Unpublished Data, Northrop Aircraft Division.
13. Northrop IR&D Program 84-R-1098.
14. Unpublished Data from the Premium Casting Task Force, AFS
Committe 2D.
15. Polmear, I.J., "Development of an Experimental Wrought
Aluminum Alloy for use at Elevated Temperatures," V. 1, pp.
664-665, published in Aluminum Alloys, Their Physical and
Mechanical Properties, proceedings of an International
Conference at University of VA, June, 1986.
16. Scarich, G.V. and Wu, K.C. "Hot Isostatic Pressing of
Aluminum Castings," Final Report, Contract N00019-80-C-0276,
February, 1986.
70